U.S. patent number 7,706,862 [Application Number 10/926,556] was granted by the patent office on 2010-04-27 for detecting human cancer through spectral optical imaging using key water absorption wavelengths.
This patent grant is currently assigned to Research Foundation of the City University of New York. Invention is credited to Robert R. Alfano, Jamal H. Ali, Wubao Wang, Manuel Zevallos.
United States Patent |
7,706,862 |
Alfano , et al. |
April 27, 2010 |
Detecting human cancer through spectral optical imaging using key
water absorption wavelengths
Abstract
Spectral optical imaging at one or more key water absorption
fingerprint wavelengths measures the difference in water content
between a region of cancerous or precancerous tissue and a region
of normal tissue. Water content is an important diagnostic
parameter because cancerous and precancerous tissues have different
water content than normal tissues. Key water absorption wavelengths
include at least one of 980 nanometers (nm), 1195 nm, 1456 nm, 1944
nm, 2880 nm to 3360 nm, and 4720 nm. In the range of 400 nm to 6000
nm, one or more points of negligible water absorption are used as
reference points for a comparison with one or more key neighboring
water absorption wavelengths. Different images are generated using
at least two wavelengths, including a water absorption wavelength
and a negligible water absorption wavelength, to yield diagnostic
information relevant for classifying a tissue region as cancerous,
precancerous, or normal. The results of this comparison can be used
to identify regions of cancerous tissue in organs such as the
breast, cervix and prostate.
Inventors: |
Alfano; Robert R. (Bronx,
NY), Ali; Jamal H. (Brooklyn, NY), Wang; Wubao
(Flushing, NY), Zevallos; Manuel (Woodhaven, NY) |
Assignee: |
Research Foundation of the City
University of New York (New York, NY)
|
Family
ID: |
46302669 |
Appl.
No.: |
10/926,556 |
Filed: |
August 26, 2004 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20050240107 A1 |
Oct 27, 2005 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
10825742 |
Apr 16, 2004 |
|
|
|
|
60463352 |
Apr 17, 2003 |
|
|
|
|
Current U.S.
Class: |
600/473; 600/477;
600/475; 600/310 |
Current CPC
Class: |
A61B
5/0059 (20130101); A61B 5/415 (20130101); A61B
5/4381 (20130101); A61B 5/7264 (20130101) |
Current International
Class: |
A61B
6/00 (20060101); A61B 5/00 (20060101) |
Field of
Search: |
;600/477,473,475,476,310 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Ion-Christian Kiricuta Jr. et al., "Tissue Water Content and
Nuclear Magnetic Resonance in Normal and Tumor Tissues", May 1975,
Cancer Research, vol. 35, pp. 1164-1167. cited by examiner .
J. Rosai, Ackerman's surgical pathology, vol. 2, Mosby
Incorporated, CA (1998). The Section of "Carcinoma" in Chapter 18,
"Male reproductive system and prostate and seminal vesicles", pp.
931-939. cited by other .
Dudley Williams, "Frequency assignments in infra-red spectrum of
water, "Nature" vol. 210, 194-195 (1966). cited by other .
C.H. Liu et al., "Raman, fluorescence and time-resolved light
scatterings as optical diagnostic techniques to separate diseased
and normal biomedical media", J. Photochem. Photobiol. B: Biol.,
vol. 16, 187-209 (1992). cited by other .
W. B. Wang et al., "Spectral polarization imaging of human prostate
tissues", Proceedings of SPIE, vol. 3917, 75-78 (2000). cited by
other.
|
Primary Examiner: Winakur; Eric F
Assistant Examiner: Fernandez; Katherine L
Attorney, Agent or Firm: Cohen Pontani Lieberman &
Pavane LLP
Parent Case Text
CROSS REFERENCE TO RELATED APPLICATION
This application is a continuation-in-part of U.S. patent
application Ser. No. 10/825,742 which was filed with the U.S.
Patent and Trademark Office on Apr. 16, 2004 now abandoned. This
application claims priority from U.S. Provisional Patent
Application Ser. No. 60/463,352 which was filed on Apr. 17, 2003.
The contents of the patent application and provisional patent
application are incorporated herein in their entirety by reference.
Claims
We claim:
1. A minimally invasive method for enabling detection of cancerous
tissues, the method comprising the steps of (a) performing spectral
optical imaging of a tissue substantially at one or more peak water
absorption wavelengths to generate a water absorption image; (b)
performing spectral optical imaging of the tissue at one or more
wavelengths of low or negligible water absorption to generate a
reference water absorption image; and comparing the generated water
absorption and reference water absorption images so as to identify
any substantial difference in water content between a first region
of the tissue and a second region of the tissue, such that changes
in water content in normal and cancerous tissues at a same water
absorption peak wavelength become detected, wherein steps (a) and
(b) are performed simultaneously or successively in any order.
2. The method of claim 1 wherein the one or more wavelengths of
lower or negligible water absorption include at least one of 4500
nm, 2230 nm, 1700 nm, and 1300 nm.
3. The method of claim 1 further including the step of generating a
difference image from the water absorption image and the reference
water absorption image.
4. The method of claim 1 wherein steps (a) and (b) are used to
diagnose one or more regions of cancerous tissue in a human
prostate by using at least one of: (i) one or more water absorption
peaks at 1195 nm for deep prostate cancer detection, and (ii) one
or more water absorption peaks at, 1944 nm, 2880-3600 nm, and 4720
nm for surface and subsurface prostate cancer detection or
pathology of thin slices of tissues.
5. The method of claim 1 wherein steps (a) and (b) are used to
diagnose one or more regions of cancerous tissue in at least one of
skin, a cervix, a human breast, and other human organs.
6. A minimally invasive method for enabling detection of tissue in
cancerous or precancerous tissues, the method comprising the steps
of: (a) performing spectral optical imaging of a tissue
substantially at one or more peak water absorption wavelengths
including at least one of 1195 nm, 1944 nm, 2880 nm to 3360 nm, and
4720 nm, to generate a water absorption image so as to enable an
identification of any regions of the tissue in terms of the water
content; (b) performing spectral optical imaging of the tissue at
one or more wavelengths of low or negligible water absorption
including at least one of 4500 nm, 2230 nm, 1700 nm, and 1300 nm,
to generate a reference water absorption image; wherein steps (a)
and (b) are performed simultaneously or successively in any order
to enable a comparison of the generated water absorption and
reference water absorption images so as to identify any substantial
difference in water content between a first region of the tissue
and a second region of the tissue.
7. The method of claim 6 further including the step of generating a
difference image from the water absorption image and the reference
water absorption image.
8. The method of claim 6 wherein steps (a) and (b) are used to
diagnose one or more regions of cancerous tissue in a human
prostate by using at least one of (i) one or more water absorption
peaks at and 1195 nm for deep prostate cancer detection, and (ii)
one or more water absorption peaks at, 1944 nm, 2880-3600 nm, and
4720 nm for surface and subsurface prostate cancer detection or
pathology of thin slices of tissues.
9. The method of claim 6 wherein steps (a) and (b) are used to
diagnose one or more regions of cancerous tissue in at least one of
skin, a cervix, a human breast, and other human organs.
10. A spectral optical imaging system comprising: a source of
infrared illumination; first and second polarizers; first and
second wideband filters; and a charge-coupled device (CCD) camera,
wherein the source is equipped to illuminate a tissue to be
diagnosed through the first wideband filter and the first
polarizer, the CCD camera is equipped to receive at least one of
transmitted light and back-scattered light from the tissue through
the second wideband filter and second polarizer, the first and
second wideband filters include a selection mechanism enabling
selection of at least one water absorption wavelength and at least
one reference water absorption wavelength, the water absorption
wavelength including at least one of 1195 nm, 1944 nm, 2880-3600
nm, and 4720 nm, to generate a water absorption image, and the
reference water absorption wavelength including at least one
infrared wavelength that provides negligible water absorption, the
at least one infrared wavelength that provides negligible water
absorption including at least one of 4500 nm, 2230 nm, 1700 nm,
1300 nm, 1000 nm, and 800 nm, to generate a reference water
absorption image; and a processing mechanism configured to compare
the generated water absorption and reference water absorption
images.
11. The spectral optical imaging system of claim 10, wherein said
system performs minimally invasive detection of cancerous tissues
by: (a) the CCD camera performing spectral optical imaging of a
tissue substantially at one or more peak water absorption
wavelengths by adjusting the first and second wideband filters to
pass electromagnetic energy at least one of, 1195 nm, 1944 nm, 2880
nm to 3360 rim, and 4720 nm, to generate a water absorption image
so as to enable an identification of any regions of the tissue
which have different water content relative to other regions; (b)
the CCD camera performing spectral optical imaging of the tissue at
one or more wavelengths of low or negligible water absorption by
adjusting the first and second wideband filters to pass
electromagnetic energy at the one or more low or negligible water
absorption wavelengths, the one or more wavelengths of, negligible
water absorption include at least one of 4500 nm, 2230 nm, 1700 nm,
1300 nm, 1000 nm, and 800 to generate a reference water absorption
image so as to enable an identification of any regions of the
tissue which have a different water content relative to other
regions; wherein the CCD camera generates the reference image and
the water absorption image simultaneously or successively in any
order, thereby enabling a comparison of the reference water
absorption image and the water absorption image to identify any
substantial difference in water content between a first region of
the tissue and a second region of the tissue.
12. The spectral optical imaging system of claim 11 wherein the one
or more wavelengths of low or negligible water absorption include
at least one of 4500 nm, 2230 nm, 1700 nm, and 1300 nm.
13. The spectral optical imaging system of claim 11 wherein the
processing mechanism includes a graphical processing mechanism for
generating a difference image from the water absorption image and
the reference water absorption image on a pixel-by-pixel basis.
14. The spectral optical imaging system of claim 11 wherein the
reference water absorption image and the water absorption image are
used to diagnose one or more regions of cancerous tissue in a human
prostate by using at least one of: (i) one or more water absorption
peaks at 1195 nm for deep prostate cancer detection, and (ii) one
or more water absorption peaks at 1944 nm, 2880-3600 nm, and 4720
nm for surface and subsurface prostate cancer detection; and
comparing one or more images generated using one or more water
absorption peaks with one or more images generated at wavelengths
having no or negligible water absorption.
15. The spectral optical imaging system of claim 11 wherein the
processing mechanism is configured to use the reference water
absorption image and the water absorption image are used to
diagnose one or more regions of cancerous tissue in at least one of
skin, a human breast, a cervix, and other human organs.
16. The spectral optical imaging system of claim 11 wherein the
source is an LED (light emitting diode) or white light source, the
system further comprising a coupling mechanism for coupling the
source to a tissue through an optical subsystem including at least
one of a filter, a lens, a mirror, a beam splitter, a polarizer,
optical fiber, a CCD detector, and a CMOS detector.
17. The spectral optical imaging system of claim 11 wherein the CCD
camera is a sensitive red visible to mid-IR CCD or CMOS camera
system.
18. The spectral optical imaging system of claim 11 further
comprising a computerized imaging system coupled to the CCD camera,
the computerized imaging system including a processing mechanism
for executing data collection software and for posting images to a
display screen.
19. The spectral optical imaging system of claim 11 wherein an
optical fiber probe is inserted rectally to provide rectal
illumination and collect the reference water absorption and water
absorption images to detect prostate cancer.
20. The spectral optical imaging system of claim 11, wherein the
reference water absorption image and the water absorption image
permit a diagnosis of one or more regions of cancerous tissue in a
human prostate by using at least one of: (i) one or more water
absorption peaks at 1195 nm for deep prostate cancer detection, and
(ii) one or more water absorption peaks at 1944 nm, 2880-3600 nm,
and 4720 nm for surface and subsurface prostate cancer detection;
and comparing one or more images generated using one or more water
absorption peaks with one or more images generated at wavelengths
having no or negligible water absorption.
21. The spectral optical imaging system of claim 10 wherein the
processing mechanism includes a graphical processing mechanism for
subtracting the water absorption images from the reference water
absorption images so as to enable a correlation of a tissue to be
diagnosed with any one of three states including normal, benign,
and cancerous tissues, wherein the graphical processing mechanism
is programmed to perform the subtracting such that:
.+-..function..lamda..-+..function..lamda..DELTA..times..times..times..ti-
mes..times..times..times..times..times..times..times..times..times..times.-
.times..times..function..lamda..function..lamda..times..times..times..time-
s..times..times..times..times..times..times. ##EQU00009## where
.lamda..sub.W represents one or more water absorption wavelengths,
.lamda..sub.NW represents one or more reference wavelengths having
no or negligible water absorption, and A is an intensity difference
between the water absorption image and the reference water
absorption image.
22. The spectral optical imaging system of claim 10 further
including a configuration adjustment mechanism for providing each
of the water absorption image and the reference water absorption
image in a parallel geometry and a perpendicular geometry, wherein
the parallel and perpendicular geometries are determined with
reference to orientation of the CCD camera, so as to permit a
determination of polarization dependency for the water absorption
image and the reference water absorption image.
23. The spectral optical imaging system of claim 10, wherein the
system performs a minimally invasive detection of cancerous tissues
by: (a) the CCD camera performing spectral optical imaging of a
tissue substantially at one or more key water absorption
wavelengths by adjusting the first and second wideband filters to
pass electromagnetic energy at least one of 1195 nm, 1944 nm, 2880
nm to 3360 nm, and 4720 nm, to generate a water absorption image so
as to enable an identification of any regions of the tissue which
have at least one of: (i) a lower water content and (ii) a higher
water content, relative to other regions; (b) the CCD camera
performing spectral optical imaging of the tissue at one or more
wavelengths of low or negligible water absorption by adjusting the
first and second wideband filters to pass electromagnetic energy at
one or more low or negligible water absorption wavelengths, the one
or more wavelengths of negligible water absorption including at
least one of 4500 nm, 2230 nm, 1700 nm, 1300 nm, 1000 nm, and 800
nm, to generate a reference water absorption image so as to enable
an identification of any regions of the tissue which have at least
one of: (i) a lower water content and (ii) a higher water content,
relative to other regions; wherein the CCD camera generates the
reference water absorption image and the water absorption image
simultaneously or successively in any order to perform a comparison
of the reference water absorption image and the water absorption
image so as to identify any substantial difference in water content
between a first region of the tissue and a second region of the
tissue.
24. A minimally invasive method for enabling detection of cancerous
tissues, the method comprising the steps of: (a) performing
spectral optical imaging of a tissue substantially at one or more
peak water absorption wavelengths to generate a water absorption
image so as to enable an identification of any regions of the
tissue which have at least one of: (i) less water content, and (ii)
more water content, relative to other regions; (b) performing
spectral optical imaging of the tissue at one or more wavelengths
of low or negligible water absorption to generate a reference water
absorption image so as to enable an identification of any regions
of the tissue which have at least one of: (i) a lower water
content, and (ii) a higher water content, relative to other
regions; wherein steps (a) and (b) are performed simultaneously or
successively in any order to compare the generated water absorption
image and reference water absorption images so as to identity any
substantial difference in water content between a first region of
the tissue and a second region of the tissue at a same water
absorption peak wavelength, such that, if a first region of tissue
has a substantially lower water content than a second region of
tissue, the first region of tissue is diagnosed as a cancerous or
precancerous tissue region in an early stage of cancer and if the
first region of tissue has a substantially higher water content
than a second region of tissue, then the first region of tissue is
diagnosed as a cancerous or precancerous region in a later stage of
cancer.
25. The method of claim 24, wherein the tissue is breast
tissue.
26. A minimally invasive method for enabling detection of cancerous
prostate tissues, the method comprising the steps of: (a)
performing spectral optical imaging of a tissue substantially at
one or more key water absorption wavelengths including at least one
of 1195 nm, 1944 nm, 2880 nm to 3360 nm, and 4720 nm, to generate a
water absorption image so as to enable an identification of any
regions of the tissue which have at least one of: (i) less water
content, and (ii) more water content, relative to other regions;
(b) performing spectral optical imaging of the tissue at one or
more wavelengths of low or negligible water absorption including at
least one of 4500 nm, 2230 nm, 1700 nm, 1300 nm, 1000 nm, and 800
nm, to generate a reference water absorption image so as to enable
an identification of any regions of the tissue which have at least
one of: (i) lower water content, and (ii) higher water content,
relative to other regions; wherein steps (a) and (b) are performed
simultaneously or successively in any order to compare the
generated water absorption and reference water absorption images so
as to identify any substantial difference in water content between
a first region of the tissue and a second region of the tissue at a
same water absorption peak wavelength, such that, if a first region
of tissue has a substantially lower water content than a second
region of tissue, the first region of tissue is diagnosed as a
cancerous or precancerous prostate tissue region in an early stage
of cancer and if the first region of tissue has a substantially
higher water content than a second region of tissue, then the first
region of tissue is diagnosed as a cancerous or precancerous
prostate region in a later stage of cancer.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention is directed to spectral optical imaging methods and,
more specifically, to optical imaging techniques for detecting
human cancer in prostate and other tissues.
2. Description of the Related Art
Cancer is a disease that is characterized by uncontrolled cellular
growth, whereby cancer cells continue to grow and divide in an
abnormal manner. A tumor, defined as any abnormal growth of cells,
may be classified as benign or malignant. A benign tumor remains
confined or localized to a given site, whereas a malignant tumor is
capable of invading other tissues or organs. Most cancers fall into
one of three main groups: carcinomas, sarcomas, and
leukemias/lymphomas. Of these groups, the most frequently-occurring
cancers are carcinomas. Carcinomas may develop from cells that
cover the surface of the body, cells of the internal organs, and
glandular cells. Glandular cells are found, for example, in the
breast and the prostate. Sarcomas are cancers of connective tissue,
such as muscle and bone. Leukemias are cancers of the blood forming
cells and cells of the immune system.
All cells consist of two major parts: a nucleus and a cytoplasm.
The nucleus is the cell's manager. It contains the cell's genetic
material in the form of strands of deoxyribonucleic acid (DNA). The
cytoplasm, a fluid within the cell, contains proteins,
carbohydrates, lipids, and nucleic acids in a water-based solution.
A change or mutation in the expression of genes causes cancer to
occur. In molecular terms, cancer is a genetic change that occurs
within the cell. Two distinct classes of cancer-related genes have
been identified: oncogenes and tumor suppressor genes.
Lung cancer, rectum cancer, breast cancer, prostate cancer, urinary
cancer, oral cancer, brain cancer and skin cancer represent some of
the most frequently occurring cancers. For men, the most common
type of cancer is cancer of the prostate. The risk of prostate
cancer increases with age. Accordingly, early detection of cancer
plays a vital role in reducing mortality from prostate cancer.
Present-day screening methods for prostate cancer include digital
rectal examinations and prostate specific antigen (PSA) blood
tests. There are several different grades or stages of cancer, and
these may be ranked using a well-known scale that classifies
cancerous and precancerous regions into any of five Gleason Grades,
denoted as Stages 1, 2, 3, 4 and 5. Precancerous stages (denoted as
stages 1 and 2) correspond to the early stages of cancer.
In an attempt to develop less invasive diagnostic procedures,
recent efforts have been directed towards utilization of
near-infrared (NIR) optical spectroscopy for cancer and pre cancer
detection. NIR techniques, based upon an understanding of cancer at
the molecular level, represent an important step toward early
detection of cancer. The optical spectrum of a tissue sample
contains information about the biochemical composition of that
tissue. A primary objective of NIR is to distinguish molecular
bonding within cancerous tissue from molecular bonding within
normal tissue by detecting fluorescence and Raman spectra from
native molecular markers. A gene that is responsible for prostate
cancer is attached or tagged with a certain chromophore (molecular
marker), such as dye or semiconductor quantum dots, to enhance
contrast and resolution in the NIR optical spectroscopy imaging
process. The use of molecular markers could enable the imaging
process to penetrate more deeply into tissue under examination,
thereby enabling doctors and other diagnostic personnel to obtain
more information.
State-of-the-art of present techniques for detection of prostate
cancer provide limited contrast, low resolution images that do not
enable an accurate identification of cancerous tissue. For this
reason, the digital rectal examination (DRE), ultrasound imaging,
and prostate specific antigen (PSA) blood test are currently the
most commonly utilized methods for early detection of prostate
cancer. Although X-rays, ultrasound, and magnetic resonance have
also been used to detect tumors, these techniques have limited
detection capabilities and/or create safety concerns. For example,
X-rays are not well-suited for the detection of tumors less than 1
mm in size and, moreover, represent a safety hazard to the
patient.
Optical spectroscopy techniques including fluorescence, Raman
scattering and light scattering have been used to investigate
normal, benign, precancerous and malignant tissues. For example,
NIR spectral polarization imaging has been used to image foreign
objects dyed with Indocyanine Green at different depths inside
prostate tissues. Some disadvantages of fluorescence and Raman
scattering methods are a) a point-by-point evaluation cannot be
performed; b) a weak diagnostic signal is provided, relative to the
amount of elastic scattering that occurs; and (c) direct contact
with cancerous tissue must occur in order to make a diagnosis.
Elastic scattering detection examines melanin and hemoglobin
absorption by focusing on the ultraviolet (UV) and visible regions
of light. In these spectral regions, light is highly scattered,
making it difficult to detect any microstructure changes that may
occur in a tissue sample.
For the sake of computational expediency, a simplification known as
the "diffusion approximation" has been widely utilized for
describing light propagation in biological media, especially when
scattering dominates absorption and the radiant energy fluence rate
close to the source is not known. Transport theory is based upon a
radiative transfer equation. The solution of this transfer equation
in a highly absorbing medium, such as water, surrounded by the
non-absorbing tissue, can be simplified and described by the
Beer-Lambert law. Note that water absorption is stronger than
scattering at specific wavelengths. The attenuation due to
absorption is proportional to the concentration (C) of chromophores
in tissues, such as water molecules or a specific dye. The optical
path length (d) is described by:
.function..times..times.e.times..times..function. ##EQU00001##
where A is the attenuation measured in optical densities, l.sub.0
is the light intensity incident on the medium, l is the light
intensity transmitted through the medium, a is the specific
extinction coefficient of the absorbing compound in micromolars per
cm, c is the concentration of the absorbing compound in
micromolars, and d is the distance between the points where the
light enters and leaves the medium (sample thickness). The product
(ac) is known as the absorption coefficient (.mu..sub.a) of the
medium. R is the specular reflection coefficient (Fresnel
reflection) from the surface of the sample. When adding absorbing
molecules to a host turbid medium (such as tissue), the
backscattered or transmitted signal from the sample
(water/chromophore-tissue) will be less, especially when absorption
dominates.
To calculate the absorption coefficient of a tissue sample, the
transmittance (T) or optical density (O. D.,
T=I/I.sub.0(1-R)=10.sup.-O.D) of a thin specimen (such as prostate
tissue) can be measured in the ballistic region. In a very thin
specimen where multiple scattering is negligible, such that
d.ltoreq.l.sub.s(l.sub.s is the scattering length), or where
absorption is much stronger than scattering, the measured
absorption coefficient can be obtained from:
.mu..times..times..function..times..times..function. ##EQU00002##
In relatively thicker tissues, the total attenuation coefficient of
a ballistic layer (.mu..sub.t=.mu..sub.s+.mu..sub.a) is
measured.
Pursuant to Fresnel's laws of reflection, specular reflection of
incident light from a surface is a function of polarization,
incident angle, and index of refraction. In the case of unpolarized
light, the reflected radiance from a surface is written as
.function..theta..function..perp. ##EQU00003## where .theta..sub.i
is the incident angle, R.sub.II is the reflected electric field
parallel to the plane of incidence, and R.sub..perp.is the
reflected electric field perpendicular to the plane of incidence.
For normal incidence (.theta..sub.i=0), equation (3) becomes
.function. ##EQU00004## where n.sub.i is the index of the incident
medium, and n.sub.t is the index of the transmitted medium.
A linearly polarized light incident on tissue loses its
polarization as it traverses the medium for an order of transport
length l.sub.tr, where
##EQU00005## and g is an anisotropy factor. A small portion of the
incident light is backscattered by epithelial cells, such that the
backscattered light retains its polarization in this single
scattering event. The remaining light diffuses into the underlying
tissue and is depolarized by multiple scattering. The degree of
polarization is defined as:
D=(I.sub.|-I.sub..perp.)/(I.sub.|+I.sub..perp.) (5) where the
I.sub..parallel. and I.sub..perp. are the intensities for the
parallel and perpendicular components of the reflected or scattered
light from the object, respectively.
The contrast is the difference in light intensity in an object or
image, and defined as:
C=(I.sub.max-I.sub.min)/(I.sub.max+I.sub.min) (6) where the
I.sub.max and I.sub.min are the maximum and minimum intensities of
light recorded from the object, respectively.
Scattering and absorption of tissue is caused by the presence of a
cellular nucleus (.about.10 .mu.m), nuclei (.about.3 .mu.m),
mitochondria (length .about.1 .mu.m), blood cells, glogi
(complicated shapes), cytoplasm, and other tissue structures. The
size of the scatterer and the incident wavelength determine the
type of scattering that will occur. Also, the distribution of the
scatterer size is an important factor in evaluating scattering
intensity versus angle
.theta..about..lamda. ##EQU00006## The optical parameters of
tissues, such as refractive index n, scattering coefficient
.mu..sub.s, and absorption coefficient .mu..sub.a, are responsible
for the degree of light scattering in tissue.
SUMMARY OF THE INVENTION
A primary object of the invention is to provide a minimally
invasive diagnostic technique for differentiating normal tissue
from cancerous and precancerous tissue.
Another object of the present invention is to detect changes in
water content in normal and cancer tissues.
Another object of the invention is to utilize spectral optical
imaging, elastic scattering, and polarization imaging techniques to
provide images of sufficient quality so to aid in diagnosing
cancerous tissue.
Still another object of the invention is to utilize spectral
optical imaging techniques to provide reliable noninvasive
diagnosis of prostate and breast cancer.
These and other objectives of the invention are achieved by using
spectral optical imaging in the near infrared (NIR) at one or more
key water absorption wavelengths to identify any difference in
water content between a region of cancerous or precancerous tissue
and a region of normal tissue. Water content is an important
diagnostic parameter. Our work using spectral polarization imaging
and spectroscopy can measure the difference in water content
between normal and cancer tissues. Our measurements show that the
tissues in the early stages of prostate cancer have less water
content than normal tissues. Tissue regions in the later stages of
cancer have more water content than normal tissues. The key water
absorption "fingerprint" wavelengths include at least one of 980
nanometers (nm), 1195 nm, 1456 nm, 1944 nm, 2880 nm to 3360 nm, and
4720 nm. In the range of 400 nm to 6000 nm, at least one reference
wavelength of low or no water absorption--illustratively, 4500 nm,
2230 nm, 1700 nm, 1300 nm, 1000 nm, and 800 nm--is used to generate
at least one reference image for drawing a comparison with at least
one image taken at one or more key water absorption wavelengths.
The results of this comparison are used to identify regions of
cancerous tissue, illustratively in organs such as the breast and
the prostate.
Pursuant to a further embodiment of the invention, imaging at key
water absorption wavelengths of approximately at least one of 980
nm, 1195 nm, 1944 nm, 2880 nm to 3360 nm, and 4720 nm is performed
to diagnose a tissue region for prostate, breast, or other cancer
by observing changes in optical density (O.D.) images of the region
due to water content. A reference image is generated using at least
one non water absorption wavelength, illustratively 800 nm and 1000
nm. The reference image is compared with one or more images
generated at the key water absorption wavelengths on a
pixel-by-pixel basis to generate a difference image. The difference
image (such as between 980 nm and 800 nm) is simplified by:
I.sub.980(x,y)-I.sub.800(x,y)=.DELTA.I, where I represents the
intensity of each pixel (x and y) in the image and .DELTA.I
represents the image difference between the two chosen wavelengths
(800 nm and 980 nm in this example) at substantially the same pixel
location.
BRIEF DESCRIPTION OF THE DRAWINGS
In the drawings:
FIG. 1 is a bar graph showing the relative water content of normal,
precancerous, and cancerous tissues for each of a plurality of
Gleason stages.
FIG. 2 is a photograph illustrating a typical specimen of human
prostate tissue.
FIG. 3 is a functional hardware block diagram of a spectral
polarization imaging system for use with the techniques of the
present invention.
FIG. 4 is a graph showing the optical density of normal prostate
tissue as a function of wavelength.
FIG. 5 is a graph comparing the optical densities of normal
prostate tissue, cancerous prostate tissue (300 .mu.m ), and water
as a function of wavelength, with a graphical inset showing the
optical density of water (1 cm thickness) throughout a spectral
range from 400 nm to 1300 nm.
FIG. 6 is a graph comparing the optical densities of normal
prostate tissue and cancerous prostate tissue as a function of
wavelength.
FIG. 7 is a graph showing curve fitting for optical density as a
function of wavelength for normal tissue.
FIG. 8 is a graph showing curve fitting for optical density as a
function of wavelength for the cancerous tissue.
FIG. 9 shows transmission images of the specimen of FIG. 2 at
several wavelengths along a parallel plane and a perpendicular
plane.
FIG. 10 shows backscattering images of the specimen of FIG. 2 at
several wavelengths along a parallel plane and a perpendicular
plane.
FIG. 11 is a graph showing optical intensity distribution at 700 nm
and 800 nm as a function of pixels for a digitized horizontal scan
from left to right at the center of the transmission images of FIG.
9.
FIG. 12 is a graph showing optical intensity distribution at 1200
nm and 1450 nm as a function of pixels for a digitized horizontal
scan from left to right at the center of the transmission images of
FIG. 9.
FIG. 13 is a layer structure of a model rectum-membrane-prostate
tissue sample made of a small dot piece of black absorber embedded
inside a larger piece of host prostate tissue in a
rectum-membrane-prostate tissue structure at a depth of 2.5 mm from
the surface of the rectum.
FIGS. 14(a) thru 14(c) show scattered light images recorded at
wavelengths of (a) 600 nm, (b) 700 nm, and (c) 800 nm,
respectively, where P and D are is the pump and detection
wavelengths, respectively.
FIG. 15 shows a schematic diagram of an optical fiber-probed NIR
polarization imaging instrument for prostate cancer detection
through rectum.
TABLE 1 sets forth calculated extinction coefficients (.mu..sub.t),
optical densities (OD), and transmission (T) for human prostate
normal tissue (N), human prostate cancerous tissue (C), and water
(W).
TABLE 2 sets forth the degree of polarization of the normal and
cancerous cells shown in FIG. 9 as a function of wavelength.
DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS
The most abundant constituent of tissue is water. Approximately 78%
of the human body is water, with the effect that water is a
universal solvent for most biological tissues. At the molecular
level, one interesting characteristic of water is that it is a
polar substance, such that one portion of the molecule carries a
negative charge and another portion carries a positive charge. This
property is important in the context of cancer diagnosis. Cancerous
tissues have a lower degree of organization and different water
content relative to normal tissues. In cells, water is essential
for converting mechanical energy generated by contractile proteins
into chemical energy that is useful for various metabolic
processes. Regulating water volume within a living cell,
contractile proteins mechanically control ion selectivity, ion
accumulation, and electron transport in mitochondria. When the
availability of water in the cell is increased, this causes a
corresponding increase in the dielectric constant of the medium,
signifying that the energy needed in ion exchange is minimized when
intracellular water is abundant.
In men, prostate cancer has a high incidence of occurrence as well
as a high mortality rate. Every year, nearly 180,000 new prostate
cancer cases are diagnosed, and about 37,000 deaths annually are
caused by prostate cancers in U.S. Current methods for monitoring
the prostate include a prostate specific antigen (PSA) blood test,
a digital rectal examination (DRE), and transrectal ultrasound
(TRUS). The PSA tests and DRE exams frequently result in false
positives. The positive predictive value of TRUS is low, and its
spatial resolution is poor. When the PSA level is elevated or the
DRE abnormal, there is a one-in-three chance that cancer is
present. Cancer can only be confirmed by a needle biopsy of the
prostate. In the biopsy, a number of cores of prostate tissue are
taken with a thin needle guided into selected regions of the
prostate with an ultrasound probe. Since ultrasound imaging has
poor spatial resolution and limited accuracy, and needle biopsy is
invasive, better approaches are needed to provide high resolution
images in a noninvasive way, so as to enable detection of prostate
tumors at an early stage. There are five different grades or stages
of cancer, oftentimes referred to as stages 1, 2, 3, 4 and 5.
Stages 1 and 2 are the early stages of cancer, and are used to
denote precancerous tissues.
Extensive research has focused on nuclear magnetic resonance
spectroscopy (NMR) techniques. Basically, NMR detects signals
generated by the nuclear spins of protons, such as the protons (H+
ions) of water. NMR spectroscopy has been used to study water in
muscle tissue. It has been shown that the water spectrum of rat or
mouse skeletal muscle is broader than that of pure water, due to
the higher order phases of water. This restriction is due to
interactions between water molecules and cellular or other
macromolecules.
The spectral properties of light propagating in tissues can be used
to evaluate the cancerous state of tissues. Under light
illumination, normal and cancerous prostate tissues absorb and emit
different light, each with unique fingerprint spectra. About 95% of
prostate cancers are categorized as adenocarcinoma, including large
duct cell, endometrial type (endometrioid), mixed edenocarcinoma,
mucinous, adenosquamous and adenoid cystic carcinoma. As shown in
FIG. 1, these cancers contain less water at the early stages
(Gleason stages 1 and 2) and, therefore, feel harder and more
condensed than normal tissue.
We have studied differences in absorption, emission and scattering
between normal and cancerous tissues, and have developed tissue
scattering light imaging, tissue emission light imaging and
contrast agent emission light imaging techniques, which
significantly enhance the visibility of an object hidden within
tissues from several millimeters to a few centimeters using 700 to
1000 nm radiation.
The interaction between light and tissue is wavelength dependent.
Well-defined wavelengths are absorbed by chromophores, such as
proteins, water, and adipose that are naturally present in tissue.
Water is involved in various chemical reactions that are activated
by light. Bonding of water molecules to other components in tissues
give rise to a 3434 cm.sup.-1 absorption peak, which is essentially
a shift in the --OH absorption peak to 3434 cm.sup.-1 due to
formation of H(hydrogen) bonds between water and tissue. The
development of NIR and mid-IR spectroscopy techniques to detect the
presence of water in tissues offers a safe, non-invasive monitoring
of the state of tissue, representing a landmark achievement in the
field of medicine. The magnitude of the aforementioned absorption
is directly related to the concentration of water in a biological
sample. The monitoring of water concentration may be advantageously
exploited to determine the state of tissue, thus aiding in the
diagnosis of cancerous, precancerous, and normal tissues.
Scattered intensity is related to R, where R is the specular
reflection coefficient for Fresnel reflection from the surface of a
tissue sample, as was previously discussed in connection with
equation (4). The index of refraction, n, of the tissue is
substantially in the range of (1.33.ltoreq.n.ltoreq.1.5), where n
takes the minimum value in this range when the content of water in
tissue is maximum (100% water in tissue), and n takes the maximum
value in this range when the content of water is minimum (0% water
in tissue). The refractive index of a tissue is proportional to its
water content, and is given by: n.apprxeq.1.5-(1.5-1.33)V (7) where
V is the volume fraction of water. The index of refraction of
cancerous tissue is higher than that of normal tissue at early
stages, since the content of water in the cancerous cells is less
than that of the normal cells. Accordingly, the backscattered light
from cancerous cells is expected to be larger than that from normal
cells. For advanced cancerous stages, the water increases give rise
to lower indices of refraction. The backscattered light in such
cases will be less.
The nuclei of cancerous cells, as well as those of normal cells,
are considered to be much larger than the wavelength of incident
light. Therefore, these nuclei obey Mie scattering, resulting in a
strong forward scattering of incident light. Since the nuclei of
cancer cells are larger than the nuclei of normal cells, the
forward scattering intensity of cancerous cells is of greater
magnitude than that of normal cells. So, the overall light
transmission of cancerous cells is greater than that of normal
cells.
The techniques of the present invention are based upon the overall
concept that, in order to detect regions of cancerous tissue, one
must realize that the amount of water contained within normal
tissues differs from the amount of water contained within
neoplastic tissues. There is a lack of water in neoplastic tissues
relative to the water content of normal tissues during the early
stages. Visible to mid-infrared (mid-IR) absorption is directly
related to the concentration of water in a biological sample.
Monitoring the concentration of water enables a determination of
whether or not regions of cancerous tissue are present. Optical
images can be performed at pairs of wavelengths: one at an
absorption wavelength of H.sub.2O and another at an off-absorption
wavelength of H.sub.2O. Difference images generated from the
absorption wavelength and off-absorption wavelength images can be
used to locate tissues in different stages of cancer.
A critical marker for locating cancerous regions in human prostate,
breast, and other tissues is the amount of water detected in these
tissues by means of transmission and backscattering of specific key
wavelengths of visible to mid-infrared (IR) light using
polarization imaging techniques. Optical interaction in the tissue
due to intermolecular bonding by the --OH portion of water
molecules is detected by visible to mid-IR spectroscopy, thus
distinguishing localized regions of low water concentration in
cancerous and precancerous tissues from other regions of normal
water concentration that occur in normal tissues. By using water as
a key marker to differentiate normal and cancerous tissue regions,
significant progress can be made towards the development of optical
non-invasive medical diagnosis in cancer research.
The techniques of the present invention are based upon a
realization that differences in light absorption are attributable
to --H and --OH bonding in tissue. In turn, the extent of --H and
--OH bonding is directly related to the water content of the tissue
under test. Typically, there is a reduction of water content in
cancerous and precancerous tissue regions relative to that of
normal and benign tissues in early stages, while the reverse is
true in later cancerous stages. The difference in light absorption,
resulting from the differing amounts of water present in normal and
cancerous tissues, can be used to diagnose a tissue region as
cancerous, precancerous, or normal.
EXPERIMENTAL METHODS
Prostate tissue specimens were obtained from the National Disease
Research Institute (NDRI) under IRB at the City Colleges of New
York (CCNY). A photograph of a typical sample of human prostate
tissue is shown in FIG. 2. This photograph was taken using a
conventional digital camera. Sample thickness is about 330 .mu.m,
and the area of the sample is approximately 2.times.3 cm.sup.2.
Throughout the various drawings, samples are arranged, if possible,
such that the right hand side of the specimen contains
predominately cancerous tissue, while the left hand side contains
predominately normal tissue.
The light absorption spectra of the normal prostate tissue, the
cancerous prostate tissue, and water were measured using a
Perkin-Elmer Lambda 9 UV/VIS/NIR Spectrophotometer with
accompanying software. Wavelengths in the approximate range of 400
nm and 25 .mu.m were utilized for this measurement process.
Images of scattered light from human prostate samples were measured
using a spectral polarization imaging system 200 as shown in FIG.
3. The system is capable of providing images using transmission
geometry as well as a back-scattering geometry. When the
transmission geometry was employed for imaging measurements, a
white light beam 223 having a diameter of approximately 2 cm was
used to illuminate a sample 213. Pursuant to transmission geometry,
the sample was positioned between the white light beam and a
charge-coupled-device (CCD) camera 219. On the other hand, when the
back-scattering geometry was used for imaging measurements, white
light beam 223 was used to illuminate sample 213 from a direction
such that some of the light scattered by sample 213 would reach CCD
camera 219.
In both the transmission geometry and the back-scattering geometry,
wideband filters (WBF) 205, 209 having a selectable bandpass for
admitting any one of several different wavelengths, such as 700 nm,
800 nm, 1200 nm, and 1450 nm, were used to select the desirable
spectral range of the illumination and the detected light. A first
polarizer (P.sub.1) 207 was located in the incident light beam
pathway to obtain a linearly polarized illumination light. A second
polarizer (P.sub.2) 211 was positioned in front of CCD camera 219
for selecting polarization direction to be detected, which may be
either parallel or perpendicular relative to the orientation of
first polarizer (P.sub.1) 207. In the visible and NIR range
(600-900 nm), CCD camera 219 was implemented using a cooled CCD
Silicon camera (Photomatrix CH250) equipped with a zoom lens of
50-mm focal length to record images in the transmission and
backscattering geometries. In the range of 1200 nm to 1450 nm, CCD
camera 219 was implemented using an InGaAs NIR CCD camera. The
images formed in CCD camera 219 will be recorded by a computer 211
through an electronic control unit 217.
EXPERIMENTAL RESULTS
FIG. 4 is a graph showing the optical density of normal prostate
tissue as a function of wavelength. Wavelengths in the range of 400
to 25,000 nm were tested. FIG. 5 is a graph comparing the
absorption spectra of normal prostate tissue (330 .mu.m thickness),
cancerous prostate tissue (330 .mu.m thickness), and water (200
.mu.m thickness) for wavelengths between 400 and 2400 nm. In the
graphs of FIGS. 4 and 5, the extent of absorption at various
frequencies is referred to as "optical density" (O.D.). The
absorption of 1 cm thickness of water is inserted in FIG. 5. For
pure water that is not associated with other molecules, the
fingerprints of absorption in the spectral range of 400-2400 nm are
980 nm (very weak), 1195 nm (weak), 1444 nm (strong), and 1930 nm
(very strong). Although these absorption fingerprints may shift
slightly in wavelength when the water molecule is associated with
tissue, these fingerprints can nonetheless be utilized as guides in
detecting the water content of tissue.
The absorption of water between 400 nm-800 nm is almost flat. The
absorption of water in the region of visible light is very small
compared to that of longer wavelengths, such as 1444 nm and 1930
nm. The absorption at 1444 nm is due to the first overtone of --OH
stretching in the water molecule. It is well known that the
absorption of the stretching vibration of the O--H bond in a
nonassociated (free) alcholic or phenolic hydroxyl group produces a
strong band at 3600 to 3650 cm.sup.-1 (2.78-2.74 .mu.m,
respectively) in the fundamental region and near 7100 cm.sup.-1
(1.41 .mu.m) in the first overtone. Reference points with low
and/or no absorptions at 1700 nm, 1300 nm, 1000 nm and 800 nm are
used to compare with water strong absorption bands at 1930 nm, 1440
nm, 1195 nm, and 980 nm. The graphical inset at the upper right
hand corner of FIG. 5 (when FIG. 5 is oriented such that the
wording appears upright) shows the optical density at spectral
range from 400 nm to 1400 nm with 1 cm thickness of water. The
measurements was done with 1 cm thickness of water indicating that
the cancerous tissue grows in the deep prostate even a few
centimeters from the surface can be determined using the water
absorption peaks at 980 nm and 1195 nm. These wavelengths (such as
980 nm and 1195 nm) offer a probe of deep cancerous and
precancerous tissue detection.
It is well known that scattering is a smooth function of wavelength
while absorption is represented by distinct peaks substantially at
one or more discrete wavelengths. The optical density spectra of
cancer and normal prostate tissues shown in FIG. 5 includes
sharply-peaked absorption bands superimposed on a smoothly varying
background caused by the prostate tissue scattering some of the
incident light. It can be concluded from the optical density graph
of FIG. 5 that scattering from cancer tissue is stronger than the
scattering from normal tissue in a forward direction between
400-1300 nm. Transmission (T) is related to optical density (O.D.)
by the formula T=10.sup.-O.D., since the O.D. for normal tissues is
greater than that of the cancer tissues then the transmission of
normal tissues is less than that of cancer. This is due to two main
factors: absorption and scattering. In the 400-1300 nm region, the
signal is mainly due to scattering. The received light intensity
from cancerous tissues is larger than the received light intensity
from normal tissues in a forward direction since the O.D. of the
cancerous tissues is smaller than that of normal tissues. Images
using CCD camera 219 (FIG. 3) show more light intensity from
cancerous tissues than normal tissues in the forward direction.
This phenomenon arises from the fact that the sizes of cells and
structures in cancerous tissue are larger than those of normal
tissues. Observations confirm Mie theory: the larger the particle
size, the greater is the forward scattering. Light transmission
through cancerous tissues is greater than that for normal tissues,
as shown in the transmission mode images of FIG. 9. The forward
scattered light from cancerous tissue arrives earlier than light
that travels through normal tissue, while at large angles, normal
tissue scatters light more strongly than cancerous tissue.
The nuclei of both cancerous and normal cells are considered to be
large particles, much larger than the visible to near infrared
wavelengths employed by the imaging process. Accordingly, these
nuclei obey Mie scattering, resulting in a strong forward
scattering of light. The scattering angle .theta..sub.S can be
written in terms of scattering wavelength (.lamda.) and the size of
the scatterer (a) as
.theta..about..lamda. ##EQU00007## The sizes of structures and
cells for cancer are larger; therefore, the scattering angle
(.theta..sub.S) is small for cancerous cells, giving a larger
intensity in the forward direction. Normal cells will scatter light
at larger angles than cancerous cell tissues. For objects having
smaller scattering sizes, such as mitochondria (much smaller than
normal size), scattering in the backward direction is larger,
giving a stronger signal for scattering off small structures.
At 1456 nm and 1944 nm, absorption dominates, such that absorption
is stronger than scattering. The graphs of FIGS. 4 and 5 show
absorption of normal tissue is stronger than that of cancerous
tissue at 1456 nm and 1944 nm, which indicates that the content of
water in normal tissues is greater than that of cancer tissues. The
peaks of around 1456 nm and 1944 nm in prostate tissue are due to
water-tissue interaction, resulting in a wavelength shift toward
longer wavelengths due to the stretching frequency of a bonded OH
group (causing a shift towards the lower wave numbers). This
wavelength shift is probably caused by the higher order phases of
water and their interactions with cellular or other macromolecules
in prostate tissues.
The calculated extinction coefficients of water at different
wavelengths are given in Table 1. The extinction coefficient of
water at 700 nm is approximately 0.433 cm.sup.-1 (the attenuation
length about 2.31 cm), 1.29 cm.sup.-1 at 1200 nm, and 9.7 cm.sup.-1
at 1450 nm. The attenuating length at 1450 is approximately 7.5
times shorter than that at 1200 nm and approximately 22 times
shorter than 700 nm in water. To reduce the effect of scattering in
the profile shown in FIG. 5, a smooth fitted curve that reflects
the contribution of scattering is subtracted from the original
curve. The result is shown in FIG. 6. The absorption fingerprints
in the visible region are 420 nm and 570 nm, which is due to the
blood in the tissue matrix (Hb and HbO.sub.2).
In cancerous tissue, the path length (equal to 1/.mu..sub.t) at
1450 nm is approximately 1.2 times shorter than at 1200 nm whereas,
in normal tissue, the path length at 1450 nm is approximately 1.3
times shorter than at 1200 nm. The total attenuation coefficient of
normal tissue is larger than that of cancerous tissue (as seen in
Table 1). The path length of normal tissue is shorter than that of
cancerous tissue. This signifies that photons traversing through
normal tissue will be absorbed or scattered at a shorter distance
than would be the case in cancerous tissue. The attenuating
length
.times..times..sigma. ##EQU00008## is inversely proportional to the
number of particles per unit volume (n) and the cross section of
the scatterer (.sigma.). Since the cross section of cancer cells
(larger nucleus) is larger than that of normal cells and the
attenuation length of normal tissues is smaller than that of
cancerous tissues (Table 1), the number of normal cell nuclei per
unit volume must be larger than that for cancerous tissues
(n.sub.n)n.sub.c).
The attenuation intensity of prostate tissues in 400-1200 nm was
fitted to C.lamda..sup.-n. In this fitting, n takes approximately
the value of 0.82 for normal tissues and 0.86 for cancerous
tissues, with different values for the C factor as shown in FIGS. 7
and 8, respectively. The n values for both normal and cancerous
tissues are close.
The scatterer size (d) of the nucleus to the wavelength (.lamda.)
(at .about.1 .mu.m) is approximately 5 times (d/.lamda..about.5) in
the normal cell and 10 times (d/.lamda..about.10) in the cancerous
cell. This is the large particle case (Mie theory), where the
scattering is stronger in the forward direction in both cases. When
n=4 (in C.lamda..sup.-n), as in the case of very small particles
(compared to the incident wavelength), this represents a scenario
where Raleigh scattering dominates. It is expected that, for larger
particles, n becomes a smaller value, so as to reduce the
scattering coefficient, as this is related to scattering
intensity.
FIG. 9 shows eight transmission images, labelled a-h, of cancerous
and normal tissue samples at 700 nm, 800 nm, 1200 nm, and 1450 nm
for parallel and perpendicular orientations of tissue. The left
piece of the specimen (predominately normal tissue) has less
transmission intensity than that on the right side (predominately
cancer) at all wavelengths (700 nm, 800 nm, 1200 nm, and 1450 nm)
as shown in FIG. 9. Similar results were obtained in normal and
cancerous human breast tissues using picosecond temporal time gated
imaging at 800 nm through the use of a Ti:sapphire pulsed laser. In
the large particle case (Mie scattering), the intensity of forward
scattering is higher than that of backscattering. Since the nuclei
of the cancer tissues are larger than that of normal tissues,
forward scattering for cancerous tissue is expected to be larger
than that of normal tissue in the forward direction. At 1200 nm,
scattering is stronger than absorption. The forward scattering
intensity from cancer tissues at 1200 nm is higher than that of
normal tissues, as shown in FIG. 5. As a result, transmission
through cancerous tissues is greater than that of normal tissues,
as shown in images c (parallel orientation) and g (perpendicular
orientation) of FIG. 9. At 1450 nm, absorption dominates (stronger
than scattering), and the absorption of normal tissue is stronger
than that of cancerous tissue, as shown in FIGS. 5 and 6. The
transmission intensity through normal tissues is weaker than that
of cancerous tissues. At the absorption peaks of water, tissue that
contains more water will absorb more incoming photons than tissue,
which contains less water. Local deviations in water concentration
within tissue will cause a differentiation in the degree of
scattering. The changes displayed in images d and h of FIG. 9
result mainly from absorption of water in tissue (first overtone of
OH stretching vibration); in addition, the forward scattering in
cancerous tissues is greater than that of normal tissues.
From the curves displayed in FIGS. 5 and 6, the absorption peak at
1450 nm is stronger than that at 1200 nm. Scattering at 1450 nm is
less than that at 1200 nm. Most of the photons at 1450 nm are
absorbed strongly by water molecules in the prostate tissues.
Photons at 1200 nm get absorbed less. Since cell nuclei are larger
than the wavelength, these nuclei predominantly scatter light is in
the forward direction. The scattered intensity is related to the
population density of the nuclei. For the perpendicular case,
depolarization is due mainly to multiple scattering events. Such
depolarization, attributable to cell size, cell shape and cell
water content, causes photons to be more depolarized in cancer
tissue since cancer is more randomized in shape and size and
includes less water content. The internal structures of the
cancerous tissues randomize the light more than in the case of
normal tissue. Normal tissue is highly ordered in water, as is
readily observed by considering the images shown in FIG. 9.
In images taken using the backscattering geometry of FIG. 3, light
scattering from cancerous tissue is stronger than that of normal
tissue. It is known that the index of refraction for cancerous
tissue is higher than that of normal tissue for early stages of
cancer (refer to equation (7) provided above). Accordingly, cancer
tissue contains less water than normal tissue and, consequently,
cancer tissue has higher index of refraction than normal tissue. As
shown in FIG. 10, one would expect that backscattering intensity
for cancerous tissue is larger than that of normal tissue, due to
the fact that cancerous tissue is denser (higher index of
refraction) than normal tissues, and due to the lower light
attenuation at water absorption wavelengths in cancerous tissue.
Moreover, smaller cellular structures, such as mitochondria, play a
major role in the backscattering geometry. As a result of the
foregoing factors, cancerous regions will appear brighter than
normal regions.
A digitized horizontal scan from left to right at the center of
transmission images a, b, c, and d displayed in FIG. 9 are shown in
FIG. 11 (700 nm and 800 nm) and FIG. 12 (1200 nm and 1450 nm). The
curves in FIGS. 11 and 12 represent the intensity distribution of
images a and b, and c and d of FIG. 9, respectively. FIG. 11 shows
that the region of cancerous tissue scatters more than the region
of normal tissue around 700 nm and 800 nm in the forward scattering
direction. The main difference between cancerous and normal tissues
in the 700 nm and 800 nm regions is attributed to scattering, since
absorption is almost identical in both cases. The images of FIG. 9
show that the cancerous region absorbs less light than the normal
region at 1450 nm and 1200 nm, due to the water content of the
tissue. Wavelengths that are not substantially absorbed by water,
such as 1700 nm, 1300 nm, 1000 nm, and 800 nm can be used to
generate reference images, so as to provide a basis of comparison
to images generated using water absorption wavelengths. Different
images at different wavelengths will provide highlights of cancer
regions for diagnoses. In addition, the forward scattering of
cancerous tissues is larger than that of normal tissues due to the
larger size of the cellular nuclei in cancerous tissue.
Accordingly, transmission through cancerous tissues is higher than
that of normal tissue, as is shown in FIG. 9. At a wavelength of
1450 nm, absorption dominates, so the primary reason for higher
transmission in cancerous regions is due to less water content in
cancerous tissue relative to regions of normal tissue, which in
turn, is related to the microscopic bonding of OH in cancerous
tissue.
A linearly polarized light incident on tissue loses its
polarization as it traverses the medium. A portion of the incident
light is backscattered by the tissue surface, retaining its
polarization in this single scattering event. The remaining light
propagating in a turbid medium, such as prostate tissue, can be
viewed as consisting of three components: ballistic, snake and
diffusive. Diffusive light is the dominant component, consisting of
multiple-scattered photons that travel the longest path and,
consequently, take the longest time to exit the sample. Ballistic
photons traverse the shortest path, retain most characteristics of
the incident photons, and carry direct information about the
interior structure of the scattering medium. Snake photons follow
ballistic photons in time and are involved in fewer scattering
events; they retain a significant amount of the initial properties
and information on structures hidden in the scattering medium.
The calculated degree of polarization (D as written in equation 5)
for normal and cancerous tissues at different wavelengths using the
data shown in FIGS. 11 and 12 is shown in Table 2. The values of D
for normal tissues are higher than that of tissues at all
wavelengths (700 nm, 800 nm, 1200 nm, and 1450 nm). This result is
due to greater randomization (abnormal growth) of cancerous cells,
whereas normal cells are more ordered. The degree of polarization
of cancerous and normal tissues increases as the wavelength
increases. The degree of polarization ratio for 1450 nm to 1200 nm
is approximately 1.1 for normal tissues and 1.7 for cancerous
tissues, which suggests that the water content of prostate tissue
affects the degree of polarization. The OH vibrational mode at 1450
nm plays an important role in both cancerous and normal tissues.
The degree of polarization for both normal and cancerous tissues at
1450 nm is due to strong absorption bonding. While at 1200 nm the
OH vibrational mode is weak and macroscopic scattering dominates,
so the shape and size play a very important role. In both cases,
the degree of polarization of cancer is less than that of normal
(D.sub.cancer<D.sub.normal).
The calculated contrasts (C as written in equation 6) between
cancer and normal tissues are 0.11 at 700 nm and 800 nm, 0.17 at
1200 nm, and 0.15 at 1450 nm. The main difference between 1200 nm
and 1450 nm contrasts is that at 1450 nm, the resulting contrast is
due to microscopic OH bonding in prostate tissue, while at 1200 nm
the difference is due to macroscopic scattering size and population
density in the prostate tissue.
The absorption spectrum and imaging measurements clearly show that
the water fingerprint absorption peaks at 980 nm, 1195 nm, 1456 nm,
1944 nm, 2880-3600 nm, and 4720 nm can be used to determine the
water contents of tissues and diagnose the cancerous tissue. Among
these wavelengths, absorption peaks at 980 nm and 1195 nm can be
used to detect deep cancerous and precancerous growing tissues a
few centimeters deep from the surface of the prostate (as shown in
the graphical inset of FIG. 5). Other wavelengths of 1456 nm, 1944
nm, 2880-3600 nm and 4720 nm can be used to detect cancerous
tissues growing on the surface and subsurface of the prostate, or
in thin sections of tissue used in pathology.
Similar to the digital rectal examination through rectum for
checking an abnormal prostate in clinical, the best way to
optically image prostate tumors is illuminating and imaging the
prostate gland through rectum. For this reason, we have imaged
objects hidden inside prostate tissues through
rectum-membrane-prostate tissues.
The sample used for the scattered light imaging measurements
consisted of a small piece of absorber (.about.1 mm) embedded
inside a large slice of prostate tissue (.about.30.times.20 mm) in
a rectum-membrane-prostate structure with a depth of .about.2.5 mm
from the surface of the rectum-membrane-prostate tissue structure
[see FIG. 13]. During the measurements, the illumination and
detection wavelengths (as an example at 600 nm, 700 nm, and 800 nm)
were synchronously changed so that the detection wavelengths were
always kept the same as that of illumination.
The scattered light images recorded at the wavelengths of 600 nm,
700 nm, and 800 nm with the detection polarization perpendicular to
that of illumination are shown in FIGS. 14(a)-14(c). It can be seen
that the object (absorber) cannot be distinguished by the 600 nm
image, but it can be clearly identified as a dark point by the 800
nm image. As the wavelength increases from 600 nm to 800 nm, the
visibility of the object improves. The wavelength dependence of the
image quality of the scattered light images can be explained by the
relative absorption spectra of the prostate and rectum tissues,
which was shown in FIG. 5. The relative absorptions of the prostate
and rectum tissues decrease when the wavelength increases from 400
nm to NIR. The short wavelength (such as 600 nm) light was absorbed
and scattered strongly by the surface and near surface layers of
the rectum-membrane-prostate tissues, and could not reach the
object deeply embedded in the host prostate tissue. In this case,
the scattered light images are formed by the light scattered only
from the surface and near surface tissue layers with almost no
contribution from the object, and therefore, the object cannot be
identified. In contrast, the larger penetration of the longer
wavelength NIR light in rectum-membrane-prostate tissues enables
them to reach the deeper object. Once the NIR light reaches the
object, the difference of scattering and absorption properties
between the foreign object and the surrounding tissues is reflected
in the image, and therefore, the foreign object can be identified
by the NIR scattering images.
Our results indicate the possibility of the spectral polarization
optical imaging technique for detecting small objects and
structural changes in prostate tissues through
rectum-membrane-prostate tissues, and the potential of imaging and
detecting prostate cancers through rectum in real time without
removing tissues using the key water absorption wavelengths.
It is clear that images at other wavelengths such as 1195 nm, 1456
nm, 1944 nm, 2880 nm, and 4720 nm are suitable to distinguish
cancer from normal prostate tissues due to the difference of
cancerous and normal prostate tissues in water concentrations as
explained earlier.
Detecting of prostate tumors using the optical rectal coherent
fiber-probed spectral polarization imaging instrument is shown in
FIG. 15. The illuminating light reaches the rectum and prostate
through an optical coherent fiber-bundle probe. The Light
back-scattered from the prostate is collected by a coherent
fiber-bundle. An image-collecting coherent fiber-bundle is coupled
into a CCD detector after the collected light passes through band
pass filters. A polarizer (linear and/or circular polarization
element) for the incident beam is coupled into a fiber. The
backscattered light from the prostate is collected by a probe using
a polarization preserving fiber. The image-collection coherent
fiber-bundle is coupled into NIR and Mid-IR CCD camera after the
collected light passes through a detecting polarizer (linear and/or
circular polarization element) called analyzer. The analyzer can be
rotated in the parallel or perpendicular polarization direction
relative to the incident polarized beam. Since the NIR and Mid-IR
polarization images have high spatial resolution and contrast, a
small prostate cancer, which cannot be detected by other methods,
may be visualized from these optical images.
* * * * *